4x speedup of people detection through Helium acceleration.
Performance boost to 13 frames per second (FPS).
Increased imaging speed, enhancing accuracy and efficiency.
4x Speedup of People Detection with Arm
Plumerai selected the Arm architecture because of its extensive reach and ecosystem. Engineers leveraged the Arm Cortex-M85 CPU with Arm Helium technology to accelerate their people-detection neural networks.
Arm Helium is a vector extension for Cortex-M-class processors that provides enhanced capabilities for executing AI and ML workloads on small, power-efficient devices. Helium includes hardware and software optimizations that help accelerate neural network model execution on Cortex-M processors, making them suitable for various applications, including smart sensors, IoT devices, wearables, and more.
Leveraging Neural Networks
The Plumerai approach to compact neural networks involves vertical integration and considering all AI layers together. In other words, they don’t treat data, models, training, inference, and hardware separately. This holistic approach is vital for efficiency.
This approach doesn’t just focus on the model architecture; that’s just a fraction of the entire process. It considers how components are intricately tied to data. Data is crucial for tiny neural networks, so gathering, curating, and correctly labeling training data is essential.
Discover the Real-World Impact of Edge AI
Explore in-depth use cases that show how edge AI is powering the next generation of IoT, solving real-world problems, driving faster decisions, and lowering costs through smarter operations right at the device level.